Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM66867
Volume:
10 |
8 |
Aug 2026
Pages:
1-5
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Computer Science & Engineering
Jawaharlal Institute of Technology, Borawan, Khargone, Madhya Pradesh, India
Computer Science & Engineering
Jawaharlal Institute of Technology, Borawan, Khargone, Madhya Pradesh, India
Crimes against children are a significant social and law-enforcement concern in India. The availability of historical crime records provides an opportunity to apply machine learning and data visualization techniques for identifying temporal and geographical patterns and supporting preventive decision-making. This paper presents a machine-learning-based framework for predictive analysis and visualization of crimes against children in India using National Crime Records Bureau (NCRB) data covering the period 2001–2023. The proposed framework consists of data preprocessing, exploratory data analysis, geographical visualization, feature preparation, model training, and predictive evaluation. Bar charts, line graphs, geographical maps, and heat maps are used to examine state-wise and temporal crime patterns. XGBoost is employed as the primary prediction algorithm because of its ability to model nonlinear relationships and complex patterns. Its performance is compared with Linear Regression and Random Forest using Accuracy, R², Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The experimental results reported in the study show that XGBoost achieved an accuracy of 97.00%, an R² score of 0.970, MSE of 3.25 × 10⁷, MAE of 4,125, and RMSE of 5,701. Random Forest achieved 94.85% accuracy and an R² score of 0.949, whereas Linear Regression achieved 89.70% accuracy and an R² score of 0.897. The visualization analysis also indicates substantial regional variation, with Uttar Pradesh, Madhya Pradesh, Maharashtra, Rajasthan, and West Bengal appearing among the states with comparatively high reported crime incidence. The proposed framework can support evidence-based analysis, resource allocation, hotspot identification, and child-protection planning.
KEYWORDS
This paper presented a machine-learning-based framework for predictive analysis and visualization of crimes against children in India using NCRB data from 2001–2023. The framework integrates preprocessing, exploratory data analysis, geographical visualization, feature engineering, and machine-learning prediction. The reported experimental results show that XGBoost achieved 97.00% accuracy and an R² score of 0.970, outperforming Random Forest and Linear Regression. It also produced the lowest reported MSE, MAE, and RMSE values. The visualization analysis demonstrates substantial temporal and geographical variation in reported crimes against children. The results suggest that predictive analytics can support evidence-based planning, hotspot monitoring, and allocation of child-protection resources. Nevertheless, the model should be considered a decision-support tool, not a replacement for law-enforcement judgment or social investigation. The interpretation of reported crime data must account for under-reporting, population differences, reporting practices, and other contextual factors.
[1] National Crime Records Bureau (NCRB), Crime in India, Government of India. [2] Chen, T., and Guestrin, C., “XGBoost: A Scalable Tree Boosting System,” Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016. [3] Breiman, L., “Random Forests,” Machine Learning, vol. 45, pp. 5–32, 2001. [4] Yu et al., research on crime-data aggregation and data-mining-based crime forecasting, as reviewed in the source dissertation. [5] Deepika K. K. and Smitha Vinod, research on crime analysis of Indian States and Union Territories using clustering and machine-learning approaches. [6] Manna and Biswas, research examining relationships between socio-economic variables and crime in India. [7] Catlett et al., research on spatiotemporal crime forecasting. [8] Wang et al., research applying ST-ResNet-based approaches to city-level crime prediction. [9] Stec and Klabjan, research incorporating external data sources for crime-count prediction. [10] Mohler et al., research addressing predictive policing and fairness considerations
Publication Date
8/20/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
1
Total views since publication
PDF Downloads
3
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...